A persistent challenge in experimental developmental psychology is determining which of many possible outcome measures best captures underlying behaviors and processes. In the looking-while-listening paradigm for studying early word comprehension, researchers have developed more than 12 distinct outcome measures, but have limited empirical basis for choosing between them. Using archival data from 15 datasets (N = 602 children, 12-60 months), we evaluated these measures' psychometric properties. We found that: (1) proportion looking, reaction time, and proportion of trials switching from distractor to target, demonstrated the strongest validity, robustness (experimental effect size), and reliability; and (2) the paradigm captures two distinct cognitive processes-detecting mismatches and confirming matches-with distractor-initial trials showing stronger developmental sensitivity. This work provides both specific recommendations for word comprehension research and a reproducible framework for evaluating measurement approaches in experimental developmental science. SUMMARY: A challenge in experimental developmental psychology is the proliferation of outcome measures for the same construct, without psychometric information to adjudicate between measures. We investigated the validity, robustness (effect size), and reliability of 12 distinct outcome measures for the looking-while-listening task, across 15 datasets collected from 602 infants. Proportion looking, reaction time, and proportion of trials switching from distractor to target were the most psychometrically sound outcome measures. Distractor-initial trials showed greater developmental sensitivity than target-initial trials. Through these two trial types, looking-while-listening captures two distinct cognitive processes: detecting mismatches and confirming matches.